The process of forming contact surface zones of tool wear in the machine vision 
system was investigated. Characteristics of the digital images surface zones 
of tool wear were obtained. Formation of the description of texture image zones
of tool wear based on a statistical approach to the computation features 
of texture with using co-occurense matrices and Markov random fields 
was implemented. Move images to the space of wavelet-transform at 
a preliminary stage of the texture images processing was offered. 
The algorithm for forming descriptions of texture images using a statistical 
approach was improved. The results of computer processing and recognition 
texture image zones of wear cutting tools are presented.